On Wed, 30 Aug 2023 22:51:31 GMT, Leonid Mesnik <[email protected]> wrote:
>> So previously we would terminate the debuggee the first time we saw any >> output on stderr, and now instead we ignore stderr. What if instead we >> process stderr the same as stdout? Would that cause a failure because the >> output is not what the test expects? What if there really is an issue >> launching the JVM that is expressed on stderr? Wouldn't we want that >> detected by this code? > > Agree, lets process stderr like stdout. This shouldn't change logic of test > but allow us at least to see stderr. I'm not seeing how calling `processDebuggeeOutput` lets us "see" stderr? If there is nothing the test needs to read from stderr then I don't see any point in processing it. IIUC this change will impact a lot of tests - correct? If so this needs a lot of testing - more than the original change that passed on the VM flags which did not catch this problem. Maybe that change should be backed out until we have a clearer idea of how to handle this extraneous output issue more broadly? ------------- PR Review Comment: https://git.openjdk.org/jdk/pull/15499#discussion_r1311001407
